Table 12 Cross-institutional generalization performance with statistical measures.

From: Novel metaheuristic optimized latent diffusion framework for automated oral disease detection in public health screening

Enhancement method

Dataset A (mean ± SD)

95% CI

Dataset B (mean ± SD)

95% CI

Dataset C (mean ± SD)

95% CI

Dataset D (mean ± SD)

95% CI

Dataset E (mean ± SD)

95% CI

Average accuracy (mean ± SD)

95% CI

ANOVA p-value

DentoSMART-LDM

96.2 ± 0.8%

[95.7, 96.7]

94.8 ± 0.9%

[94.2, 95.4]

95.1 ± 0.7%

[94.6, 95.6]

93.7 ± 1.0%

[93.1, 94.3]

94.2 ± 0.8%

[93.7, 94.7]

94.8 ± 0.9%

[94.4, 95.2]

p = 0.234

Enhanced-PSO-LDM

91.4 ± 1.1%

[90.7, 92.1]

89.6 ± 1.3%

[88.8, 90.4]

90.2 ± 1.0%

[89.6, 90.8]

88.3 ± 1.4%

[87.4, 89.2]

89.7 ± 1.2%

[88.9, 90.5]

89.8 ± 1.2%

[89.3, 90.3]

p = 0.187

GA-Diffusion

88.9 ± 1.3%

[88.1, 89.7]

87.1 ± 1.5%

[86.2, 88.0]

87.8 ± 1.2%

[87.1, 88.5]

86.2 ± 1.6%

[85.2, 87.2]

87.4 ± 1.4%

[86.5, 88.3]

87.5 ± 1.4%

[86.9, 88.1]

p = 0.156

DE-Enhancement

87.2 ± 1.4%

[86.3, 88.1]

85.4 ± 1.7%

[84.4, 86.4]

86.1 ± 1.3%

[85.3, 86.9]

84.6 ± 1.8%

[83.5, 85.7]

85.8 ± 1.5%

[84.8, 86.8]

85.8 ± 1.5%

[85.1, 86.5]

p = 0.143

Traditional enhancement

82.3 ± 1.6%

[81.4, 83.2]

80.1 ± 1.8%

[79.0, 81.2]

81.4 ± 1.5%

[80.6, 82.2]

79.7 ± 1.9%

[78.5, 80.9]

80.9 ± 1.7%

[79.9, 81.9]

80.9 ± 1.7%

[80.2, 81.6]

p = 0.124

No enhancement

76.8 ± 1.9%

[75.6, 78.0]

74.2 ± 2.2%

[72.7, 75.7]

75.6 ± 1.8%

[74.5, 76.7]

73.9 ± 2.4%

[72.2, 75.6]

75.1 ± 2.0%

[73.8, 76.4]

75.1 ± 2.1%

[74.2, 76.0]

p = 0.098

Back to article page